Re: [PATCH] xdiff: Do not enable XDL_FAST_HASH by default

2016-12-06 Thread Junio C Hamano
Jeff King writes: > This is a nice incremental step in the sense that people can still > enable it if they want to in order to time it, play with it, etc. But > given what we know, I wonder if the help text here should warn people. > > Or I guess we could move straight to dropping it entirely. >

Re: [PATCH] xdiff: Do not enable XDL_FAST_HASH by default

2016-11-30 Thread Jeff King
On Wed, Nov 30, 2016 at 11:26:43PM -0500, Anders Kaseorg wrote: > > So I suspect a better strategy in general is to just override the > > uname_* variables when cross-compiling. > > The specific case of an i386 userspace on an x86_64 kernel is important > independently of the general cross compi

[PATCH] xdiff: Do not enable XDL_FAST_HASH by default

2016-11-30 Thread Anders Kaseorg
Although XDL_FAST_HASH computes hashes slightly faster on some architectures, its collision characteristics are much worse, resulting in some pathological diffs running over 100x slower (http://public-inbox.org/git/20141222041944.ga...@peff.net/). Furthermore, it was being enabled when ‘uname -m’

[PATCH] xdiff: Do not enable XDL_FAST_HASH by default

2016-11-30 Thread Anders Kaseorg
Although XDL_FAST_HASH computes hashes slightly faster on some architectures, its collision characteristics are much worse, resulting in some pathological diffs running over 100x slower (http://public-inbox.org/git/20141222041944.ga...@peff.net/). Furthermore, it was being enabled when ‘uname -m’